Given the recent disappointments in the local multipoint distribution service (LMDS) market in North America, there is a tremendous amount of skepticism about whether the current technology and systems can evolve to enable the deployment of broadband wireless networks on a wide scale. Based on the history of hollow promises, false starts and business failures, equipment vendors to date have failed to develop cost-effective and architecturally sound fixed wireless systems. The lapse is due in part to their continued reliance on approaches based on legacy analog and former defense technology.
Certainly the shakeout of the financial markets has made more glaring the flawed business plans of some of the LMDS operators, which brings us to the current dismal state of the broadband wireless industry. But at the time when carriers initially secured their allocations of LMDS spectrum licenses (28, 31 and 38 GHz, and other high-frequency bands), it was believed that, as advertised, systems using point-to-multipoint (PMP) or shared-bandwidth architectures based on instantaneous demand would allow the widespread delivery of broadband services to the large population of underserved small- and medium-sized businesses worldwide.
As most fixed-wireless carriers soon discovered, early-generation PMP systems made success difficult given the low functionality and relatively high cost of those platforms. Not to mention, there was a high probability the systems would require displacement in a matter of one or two years because of capacity or frequency constraints. That situation proved unacceptable for carriers looking for a future-proof solution to operate for five to 10 years in accordance with the needs of their business plans.
As a result, most carriers continued to rely almost exclusively on the more traditional and lower-functionality point-to-point (PTP) systems, resulting in a high-cost rollout with a much reduced ability to cover a geographic area, given the limited architecture of PTP for widespread deployment.
As with all new technology markets, LMDS poses initial challenges in the development of commercial-quality systems that keep costs in a range where widespread adoption can occur. Today, developers can layer upon the work of early technologies and also enjoy the parallel progress of gallium-arsenide producers in greatly improving their yields to develop cost-effective RF components.
As a result, third-generation LMDS systems have been developed and are being commercially deployed by carriers. These new systems continue to use burst modems (or time-division multiple access) because of flexibility and other potential benefits. The difference is that now this is done with a much larger channel similar to the ones used by first-generation systems. The result is greatly increased capacity to develop broadband burst modems delivering speeds that reach 100 Mbits/second or more.
Third-generation approaches employ advanced modulation schemes along with TDMA to further increase bandwidth while successfully overcoming the one-modulation-scheme, one-channel problems previously encountered. In this way, the carrier is not required to triple its investment or even increase operations costs, as was the case with second-generation systems. And in providing the ultimate in flexibility, efficiency and revenue potential, the third-generation systems have implemented an on-demand technology for the allocation of transmitting and receiving bandwidth within the system, which translates into even greater cost efficiencies.
The focus of the third generation is to perform all operations on a burst basis and do it in very large channels.The technology is centered on the TDMA of broadband burst modems, allowing thousands of users to share capacity based on instantaneous demands. As opposed to the fixed packets of traditional TDMA, which is used in second-generation approaches, a further innovation in some third-generation systems is to allow TDMA to operate on a variable-length packet basis to support all protocols equally-asynchronous transfer mode, Internet Protocol, TDM, frame relay-while maintaining proper quality-of-service levels. Known as adaptive TDMA, the approach varies the packet lengths to enable carriers to deliver differentiated, bursty data services, while further allowing them to offer service-level agreements to customers.
In terms of modulation, the conflicting challenges of maximizing range or capacity, both core goals of the service providers, left many in a quandary. Third-generation systems have vaulted ahead to resolve it. With adaptive-modulation technology, systems now can maximize both bandwidth and range simultaneously.
Working in conjunction with TDMA, the system automatically selects the most efficient modulation scheme possible for each customer for each burst. Therefore, near-in customers use the more efficient modulation schemes of QAM64 and QAM16, while the range of the basestation continues to be maximized by quadrature phase-shift keying (QPSK).
The result is double the throughput over similar-size channel systems, but in real terms throughput improvements are better than fourfold when compared with second-generation systems. And for those second-generation systems where QAM16 was implemented to increase capacity on their narrow channels, third-generation systems require half as many basestations to cover the same area and still improve the throughput. By selecting the optimal modulation scheme for each user on a single channel, adaptive modulation maximizes capacity and range simultaneously. Thus, the result is that each basestation has the range of QPSK with an average capacity of QAM16. This combination substantially re-duces the number of basestations and RF carriers needed to serve a given area. With adaptive modulation, each millisecond the system selects the correct modulation scheme (QPSK, QAM16 or QAM64) for each burst based on each user's channel condition-distance to the customer, environmental and RF conditions.
One final challenge overcome by third-generation approaches was the amount of bandwidth asymmetry the system can support. As mentioned, first- and second-generation systems relied on the legacy, analog technology of frequency-division duplexing (FDD), where one channel is dedicated for upstream communications and a separate channel for downstream. Clearly in this new world of data access with constantly variable amounts of broadband bandwidth needed in either direction (to or from the network), the asymmetry in the channel varies instantaneously just as the bandwidth needs of the users do.
Further innovating on TDMA, third-generation systems use time-division duplexing (TDD) technology to allow the allocation of up and down bandwidth on a variable basis based on need. An all-digital duplexing technology known as adaptive TDD is a commercially proven technique that separates upstream and downstream traffic in time, rather than frequency, as with FDD. TDD has been successfully deployed in Personal Handyphone Service networks in Japan and throughout Europe in Digital European Cordless Telecommunications networks. Improving on narrowband TDD, adaptive TDD for LMDS allows instantaneous changes in traffic asymmetry by making real-time adjustments to upstream and downstream capacity and bursts by users.
In contrast, FDD systems that use fixed channel allocations for up and down bandwidth are inefficient in their utilization of equipment and capacity. The RF and modem parts of both transmit and receive chains operate on the same frequency but at different times. An adaptive-TDD-enabled system reuses certain elements for chains such as filters, mixers, frequency sources and synthesizers while eliminating isolation complexity completely.
As a result, carriers use a given amount of spectrum more efficiently, because adaptive TDD eliminates guardbands to separate upstream and downstream frequency traffic. Commonly as much as 200- to 300-MHz frequency separation is needed between transmit and receive frequencies for cost-effective modem designs in FDD, which results in loss of that spectrum to serve customers or increased system costs because of expensive and inefficient duplexers. Subsequently, the use of adaptive
Innovating on TDMA, 3G LMDS uses time-division duplexing technology to allow the allocation of up-and-down bandwidth on a variable basis.
TDD results in a substantial frequency savings, which is most noticeable in block allocations such as U.S. lmds A Band, with 850 MHz, or Canada's similar bands with 500 MHz. For allocations such as the U.S. lmds B Band (150 MHz), where there is not enough guardband for FDD, adaptive TDD is the only solution.
Despite the recent maladies among U.S. competitive carriers and diminishing confidence in lmds, skeptics must remember that as in all new markets, maturation of technology and systems takes time. After more than six years of innovation, the newly introduced third-generation systems have overcome the numerous technical hurdles of first- and second-generation approaches, in making systems that are flexible. The result is much lower costs of deployment and operation and a large increase in revenue potential from the efficiency of the systems and the ability to provide a new generation of bursting, broadband services.
Further, improvements in the performance and cost of RF components will no doubt aid in the mass-production of PMP lmds systems as the industry reaches a critical mass. Over the next year, wide-scale deployments of lmds and other broadband wireless networks worldwide will provide the necessary proof that fixed wireless is the only cost-effective and reliable way to deliver broadband services on a wide scale.
See related chart
See related chart